]>
Commit | Line | Data |
---|---|---|
52e95789 JH |
1 | /* |
2 | * Copyright (C) 2005 Junio C Hamano | |
f506b8e8 | 3 | * Copyright (C) 2010 Google Inc. |
52e95789 | 4 | */ |
b6c09c03 | 5 | #include "git-compat-util.h" |
52e95789 JH |
6 | #include "diff.h" |
7 | #include "diffcore.h" | |
f506b8e8 | 8 | #include "xdiff-interface.h" |
b95c5ada | 9 | #include "kwset.h" |
df6e8744 | 10 | #include "oidset.h" |
b6c09c03 | 11 | #include "pretty.h" |
b51a9c14 | 12 | #include "quote.h" |
f506b8e8 | 13 | |
61690bf4 JK |
14 | typedef int (*pickaxe_fn)(mmfile_t *one, mmfile_t *two, |
15 | struct diff_options *o, | |
16 | regex_t *regexp, kwset_t kws); | |
17 | ||
f506b8e8 JH |
18 | struct diffgrep_cb { |
19 | regex_t *regexp; | |
20 | int hit; | |
21 | }; | |
22 | ||
a8d5eb6d | 23 | static int diffgrep_consume(void *priv, char *line, unsigned long len) |
f506b8e8 JH |
24 | { |
25 | struct diffgrep_cb *data = priv; | |
26 | regmatch_t regmatch; | |
f506b8e8 JH |
27 | |
28 | if (line[0] != '+' && line[0] != '-') | |
a8d5eb6d | 29 | return 0; |
f506b8e8 | 30 | if (data->hit) |
fa59e7be ÆAB |
31 | BUG("Already matched in diffgrep_consume! Broken xdiff_emit_line_fn?"); |
32 | if (!regexec_buf(data->regexp, line + 1, len - 1, 1, | |
33 | ®match, 0)) { | |
34 | data->hit = 1; | |
35 | return 1; | |
36 | } | |
a8d5eb6d | 37 | return 0; |
f506b8e8 JH |
38 | } |
39 | ||
61690bf4 JK |
40 | static int diff_grep(mmfile_t *one, mmfile_t *two, |
41 | struct diff_options *o, | |
0ada4b9b | 42 | regex_t *regexp, kwset_t kws UNUSED) |
f506b8e8 | 43 | { |
61690bf4 JK |
44 | struct diffgrep_cb ecbdata; |
45 | xpparam_t xpp; | |
46 | xdemitconf_t xecfg; | |
fa59e7be | 47 | int ret; |
bc615898 | 48 | |
61690bf4 JK |
49 | /* |
50 | * We have both sides; need to run textual diff and see if | |
51 | * the pattern appears on added/deleted lines. | |
52 | */ | |
53 | memset(&xpp, 0, sizeof(xpp)); | |
54 | memset(&xecfg, 0, sizeof(xecfg)); | |
55 | ecbdata.regexp = regexp; | |
56 | ecbdata.hit = 0; | |
5d934600 | 57 | xecfg.flags = XDL_EMIT_NO_HUNK_HDR; |
61690bf4 JK |
58 | xecfg.ctxlen = o->context; |
59 | xecfg.interhunkctxlen = o->interhunkcontext; | |
fa59e7be ÆAB |
60 | |
61 | /* | |
62 | * An xdiff error might be our "data->hit" from above. See the | |
63 | * comment for xdiff_emit_line_fn in xdiff-interface.h | |
64 | */ | |
5d934600 | 65 | ret = xdi_diff_outf(one, two, NULL, diffgrep_consume, |
fa59e7be ÆAB |
66 | &ecbdata, &xpp, &xecfg); |
67 | if (ecbdata.hit) | |
68 | return 1; | |
69 | if (ret) | |
70 | return ret; | |
71 | return 0; | |
f506b8e8 JH |
72 | } |
73 | ||
5b0672a2 ÆAB |
74 | static unsigned int contains(mmfile_t *mf, regex_t *regexp, kwset_t kws, |
75 | unsigned int limit) | |
52e95789 | 76 | { |
102fdd2e ÆAB |
77 | unsigned int cnt = 0; |
78 | unsigned long sz = mf->size; | |
79 | const char *data = mf->ptr; | |
2002eed6 | 80 | |
d01d8c67 PB |
81 | if (regexp) { |
82 | regmatch_t regmatch; | |
83 | int flags = 0; | |
84 | ||
52e011cd | 85 | while (sz && |
b7d36ffc | 86 | !regexec_buf(regexp, data, sz, 1, ®match, flags)) { |
d01d8c67 | 87 | flags |= REG_NOTBOL; |
50fd6997 | 88 | data += regmatch.rm_eo; |
f53c5de2 | 89 | sz -= regmatch.rm_eo; |
52e011cd | 90 | if (sz && regmatch.rm_so == regmatch.rm_eo) { |
50fd6997 | 91 | data++; |
f53c5de2 SG |
92 | sz--; |
93 | } | |
2002eed6 | 94 | cnt++; |
5b0672a2 ÆAB |
95 | |
96 | if (limit && cnt == limit) | |
97 | return cnt; | |
2002eed6 | 98 | } |
d01d8c67 PB |
99 | |
100 | } else { /* Classic exact string match */ | |
ce163c79 | 101 | while (sz) { |
5d176fb6 RS |
102 | struct kwsmatch kwsm; |
103 | size_t offset = kwsexec(kws, data, sz, &kwsm); | |
b95c5ada | 104 | if (offset == -1) |
ce163c79 | 105 | break; |
e4aab504 RS |
106 | sz -= offset + kwsm.size[0]; |
107 | data += offset + kwsm.size[0]; | |
ce163c79 | 108 | cnt++; |
5b0672a2 ÆAB |
109 | |
110 | if (limit && cnt == limit) | |
111 | return cnt; | |
d01d8c67 | 112 | } |
2002eed6 JH |
113 | } |
114 | return cnt; | |
52e95789 JH |
115 | } |
116 | ||
61690bf4 | 117 | static int has_changes(mmfile_t *one, mmfile_t *two, |
0ada4b9b | 118 | struct diff_options *o UNUSED, |
5d176fb6 | 119 | regex_t *regexp, kwset_t kws) |
61690bf4 | 120 | { |
5b0672a2 ÆAB |
121 | unsigned int c1 = one ? contains(one, regexp, kws, 0) : 0; |
122 | unsigned int c2 = two ? contains(two, regexp, kws, c1 + 1) : 0; | |
5d35a953 | 123 | return c1 != c2; |
61690bf4 JK |
124 | } |
125 | ||
126 | static int pickaxe_match(struct diff_filepair *p, struct diff_options *o, | |
127 | regex_t *regexp, kwset_t kws, pickaxe_fn fn) | |
15dafaf8 | 128 | { |
bc615898 SR |
129 | struct userdiff_driver *textconv_one = NULL; |
130 | struct userdiff_driver *textconv_two = NULL; | |
ef90ab66 JK |
131 | mmfile_t mf1, mf2; |
132 | int ret; | |
133 | ||
61690bf4 JK |
134 | /* ignore unmerged */ |
135 | if (!DIFF_FILE_VALID(p->one) && !DIFF_FILE_VALID(p->two)) | |
136 | return 0; | |
137 | ||
15af58c1 SB |
138 | if (o->objfind) { |
139 | return (DIFF_FILE_VALID(p->one) && | |
140 | oidset_contains(o->objfind, &p->one->oid)) || | |
141 | (DIFF_FILE_VALID(p->two) && | |
142 | oidset_contains(o->objfind, &p->two->oid)); | |
143 | } | |
144 | ||
0d1e0e78 | 145 | if (o->flags.allow_textconv) { |
bd7ad45b NTND |
146 | textconv_one = get_textconv(o->repo, p->one); |
147 | textconv_two = get_textconv(o->repo, p->two); | |
a8f61094 | 148 | } |
bc615898 | 149 | |
ef90ab66 JK |
150 | /* |
151 | * If we have an unmodified pair, we know that the count will be the | |
152 | * same and don't even have to load the blobs. Unless textconv is in | |
153 | * play, _and_ we are using two different textconv filters (e.g., | |
154 | * because a pair is an exact rename with different textconv attributes | |
155 | * for each side, which might generate different content). | |
156 | */ | |
157 | if (textconv_one == textconv_two && diff_unmodified_pair(p)) | |
158 | return 0; | |
159 | ||
e0e7cb80 TB |
160 | if ((o->pickaxe_opts & DIFF_PICKAXE_KIND_G) && |
161 | !o->flags.text && | |
162 | ((!textconv_one && diff_filespec_is_binary(o->repo, p->one)) || | |
163 | (!textconv_two && diff_filespec_is_binary(o->repo, p->two)))) | |
164 | return 0; | |
165 | ||
6afaf807 NTND |
166 | mf1.size = fill_textconv(o->repo, textconv_one, p->one, &mf1.ptr); |
167 | mf2.size = fill_textconv(o->repo, textconv_two, p->two, &mf2.ptr); | |
ef90ab66 | 168 | |
f97fe358 | 169 | ret = fn(&mf1, &mf2, o, regexp, kws); |
ef90ab66 JK |
170 | |
171 | if (textconv_one) | |
172 | free(mf1.ptr); | |
173 | if (textconv_two) | |
174 | free(mf2.ptr); | |
175 | diff_free_filespec_data(p->one); | |
176 | diff_free_filespec_data(p->two); | |
177 | ||
178 | return ret; | |
15dafaf8 RS |
179 | } |
180 | ||
3753bd1f RS |
181 | static void pickaxe(struct diff_queue_struct *q, struct diff_options *o, |
182 | regex_t *regexp, kwset_t kws, pickaxe_fn fn) | |
183 | { | |
184 | int i; | |
185 | struct diff_queue_struct outq; | |
186 | ||
187 | DIFF_QUEUE_CLEAR(&outq); | |
188 | ||
189 | if (o->pickaxe_opts & DIFF_PICKAXE_ALL) { | |
190 | /* Showing the whole changeset if needle exists */ | |
191 | for (i = 0; i < q->nr; i++) { | |
192 | struct diff_filepair *p = q->queue[i]; | |
193 | if (pickaxe_match(p, o, regexp, kws, fn)) | |
194 | return; /* do not munge the queue */ | |
195 | } | |
196 | ||
197 | /* | |
198 | * Otherwise we will clear the whole queue by copying | |
199 | * the empty outq at the end of this function, but | |
200 | * first clear the current entries in the queue. | |
201 | */ | |
202 | for (i = 0; i < q->nr; i++) | |
203 | diff_free_filepair(q->queue[i]); | |
204 | } else { | |
205 | /* Showing only the filepairs that has the needle */ | |
206 | for (i = 0; i < q->nr; i++) { | |
207 | struct diff_filepair *p = q->queue[i]; | |
208 | if (pickaxe_match(p, o, regexp, kws, fn)) | |
209 | diff_q(&outq, p); | |
210 | else | |
211 | diff_free_filepair(p); | |
212 | } | |
213 | } | |
214 | ||
215 | free(q->queue); | |
216 | *q = outq; | |
217 | } | |
218 | ||
3d5b23a3 NTND |
219 | static void regcomp_or_die(regex_t *regex, const char *needle, int cflags) |
220 | { | |
221 | int err = regcomp(regex, needle, cflags); | |
222 | if (err) { | |
223 | /* The POSIX.2 people are surely sick */ | |
224 | char errbuf[1024]; | |
225 | regerror(err, regex, errbuf, 1024); | |
3d5b23a3 NTND |
226 | die("invalid regex: %s", errbuf); |
227 | } | |
228 | } | |
229 | ||
63b52afa | 230 | void diffcore_pickaxe(struct diff_options *o) |
52e95789 | 231 | { |
382f013b JH |
232 | const char *needle = o->pickaxe; |
233 | int opts = o->pickaxe_opts; | |
d01d8c67 | 234 | regex_t regex, *regexp = NULL; |
b95c5ada | 235 | kwset_t kws = NULL; |
03c1f14a | 236 | pickaxe_fn fn; |
52e95789 | 237 | |
2e197a75 ÆAB |
238 | if (opts & ~DIFF_PICKAXE_KIND_OBJFIND && |
239 | (!needle || !*needle)) | |
240 | BUG("should have needle under -G or -S"); | |
63b52afa | 241 | if (opts & (DIFF_PICKAXE_REGEX | DIFF_PICKAXE_KIND_G)) { |
218c45a4 | 242 | int cflags = REG_EXTENDED | REG_NEWLINE; |
c1ddc461 | 243 | if (o->pickaxe_opts & DIFF_PICKAXE_IGNORE_CASE) |
218c45a4 | 244 | cflags |= REG_ICASE; |
3d5b23a3 | 245 | regcomp_or_die(®ex, needle, cflags); |
d01d8c67 | 246 | regexp = ®ex; |
03c1f14a ÆAB |
247 | |
248 | if (opts & DIFF_PICKAXE_KIND_G) | |
249 | fn = diff_grep; | |
250 | else if (opts & DIFF_PICKAXE_REGEX) | |
251 | fn = has_changes; | |
252 | else | |
253 | /* | |
254 | * We don't need to check the combination of | |
255 | * -G and --pickaxe-regex, by the time we get | |
256 | * here diff.c has already died if they're | |
257 | * combined. See the usage tests in | |
258 | * t4209-log-pickaxe.sh. | |
259 | */ | |
260 | BUG("unreachable"); | |
15af58c1 SB |
261 | } else if (opts & DIFF_PICKAXE_KIND_S) { |
262 | if (o->pickaxe_opts & DIFF_PICKAXE_IGNORE_CASE && | |
263 | has_non_ascii(needle)) { | |
264 | struct strbuf sb = STRBUF_INIT; | |
265 | int cflags = REG_NEWLINE | REG_ICASE; | |
266 | ||
267 | basic_regex_quote_buf(&sb, needle); | |
268 | regcomp_or_die(®ex, sb.buf, cflags); | |
269 | strbuf_release(&sb); | |
270 | regexp = ®ex; | |
271 | } else { | |
272 | kws = kwsalloc(o->pickaxe_opts & DIFF_PICKAXE_IGNORE_CASE | |
273 | ? tolower_trans_tbl : NULL); | |
274 | kwsincr(kws, needle, strlen(needle)); | |
275 | kwsprep(kws); | |
276 | } | |
03c1f14a ÆAB |
277 | fn = has_changes; |
278 | } else if (opts & DIFF_PICKAXE_KIND_OBJFIND) { | |
279 | fn = NULL; | |
280 | } else { | |
281 | BUG("unknown pickaxe_opts flag"); | |
d01d8c67 PB |
282 | } |
283 | ||
03c1f14a | 284 | pickaxe(&diff_queued_diff, o, regexp, kws, fn); |
8e854b00 | 285 | |
63b52afa RS |
286 | if (regexp) |
287 | regfree(regexp); | |
15af58c1 | 288 | if (kws) |
b95c5ada | 289 | kwsfree(kws); |
52e95789 JH |
290 | return; |
291 | } |